信息熵加权的协同聚类算法的改进与优化  

Improvement and optimization of information entropy weighted collaborative clustering algorithm

作　　者：王红[1] 陈功平[1] WANG Hong;CHEN Gongping(Information and Electronic Engineering Insitute,Lu′an Vocational Technical College,Luan AnHui 237000)

机构地区：[1]六安职业技术学院信息与电子工程学院,安徽六安231300

出　　处：《宁夏师范学院学报》2020年第1期59-65,共7页Journal of Ningxia Normal University

基　　金：安徽省教育厅2019年度高校自然科学研究重点项目(KJ2019A1065);安徽省教育厅2019年度高校学科(专业)拔尖人才学术资助项目(gxbjZD74);2018年度高等学校省级质量工程项目(2018mooc340).

摘　　要：在协同聚类算法的基础上引入信息熵权重,用信息熵计算隶属度差异矩阵,极大地简化了算法过程,并通过权重贡献最终实现协同聚类.仿真实验显示,该聚类算法比标准FCM算法、HC-FCM算法的准确度更高,运行时间短,综合性能明显优于其他两种算法.Information entropy weight is introduced on the basis of collaborative clustering algorithm,and membership difference matrix is calculated through information entropy which greatly simplifies the algorithm process,and finally realizes collaborative clustering through weight contribution.Simulation results show that the clustering algorithm has higher accuracy and shorter running time than the standard FCM algorithm and HC-FCM algorithm,and the comprehensive performance of the clustering algorithm is obviously better than the other two algorithms.

关 键 词：协同聚类 协同关系 信息熵 权重系数 

分 类 号：TN915.03[电子电信—通信与信息系统]